Naming scheme

wg1.0.e == “western gesture 1, condition 0, eastern viewer culture” eg4.2.w == “eastern gesture 4 condition 2, western viewer culture”

Part One: Overlapping Metaphors!?!?

Load ‘N’ Wrangle

Western Participants

Load these bad larrys (straight from mTurk template)

Also load these buggerinos (from psytoolkit template)

Eastern Participants

Spot checks

some things should correlate, others should not. IE the questions that go together should correlate. accessible <–> open confident <–> sure conflict <–> tension dominant <–> control goal <–> worktogether many <–> members

So now we group the metaphor measures by group and facet wrap by that. EZ

Create grouped overlay of correlated questions. Violin plot of question correlations by group (for single conditions of a single gesture)

This graph shows the correlation of the grouped variables. We want them to be highly correlated with each other, and preferably not super correlated with one another. This should hold moreso for the extreme versions of the gestures.

Really could have correlation plot w Q1 on X and Q2 on Y but those don’t look great and we shouldn’t expect them to…

Correlation Matrix of questions

Get wrapped plot of correlations across conditions for same gesture

Violin plots of responses to questions

Density plot?

Regrouping data in a potentially horrifying way. Need to make sure questions make sense…

Group it all together and do it by group but you can also just do it by groups and facet wrap by metaphor measure for indiv question results.

Put it together and how do you do? Bibbity Bobbity BAM.

## Saving 7 x 5 in image

Correlations

Western Gestures, Western/Eastern Participants

Eastern Gestures, Western/Eastern Participants

To get the individual ones (per gesture, say) you can do this to position the w/e side by side.

Cool now access everything as follows: all_dat$overlays[["d0_overlay"]]: the overlayed violin plot of related questions. Illustrates density overlay aka a nice vis of correlation of questions

all_dat$correlation_matrix: the correlation matrices that visualize the above as well.

all_dat$violin_density_question: the violin plot of all question distributions across gesture conditions.

all_dat$violin_density_grouped_overlay: the violin plot of all question distributions across gesture conditions, but overlayed.

all_dat$violin_density_grouped: the violin plot of group distributions across gesture conditions.

all_dat$density_grouped: density plot of group distributions across gesture conditions.

all_dat$density_question: density plot of all question distributions across gesture conditions.

if you want to plot all of the overlays nicely you can do this:

Stats

Pretty, but what does it mean? Well, to determine whether any of these differences are significant (aka, did people interpret different things from each of the different gesture conditions, which, because we, too, are people, we know they did) we need to see what the significant differences between rankings in each gesture and condition are.

Now get those T-Tests done DID.

group cond1 cond2 p sig
openness cond2 cond1 0.05197 *
openness original cond2 0.08087 *
conflict cond2 cond1 0.00000 ***
conflict original cond2 0.00000 ***
unity cond2 cond1 0.00000 ***
unity original cond2 0.00011 ***
group cond1 cond2 p sig
conflict cond1 original 0.00001 ***
conflict cond2 cond1 0.06043 *
control cond1 original 0.09701 *
group cond1 cond2 p sig
openness cond2 cond1 0.00234 **
openness original cond2 0.00406 **
control cond2 cond1 0.01833 *
size cond2 cond1 0.01760 *
group cond1 cond2 p sig
group cond1 cond2 p sig
conflict original cond2 0.08661 *
unity cond2 cond1 0.00574 *

PART TWO: BICULTURALISM AT ITS FINEST

Wrangle these bad larries

cool now we have all that data in one place. We need to compare the western and eastern viewers across conditions.

Plot these bad larries.

Easy enough to do stats on means between cultures but baby I wanna see those VIOLIN PLOTSSSSSS.

Can only plot by one thing at a time (i.e. for a single gesture condition then visualize across metaphor measures, or for a single metaphor measure then visualize across conditions. The second doesn’t make sense though….)

## `stat_bindot()` using `bins = 30`. Pick better value with `binwidth`.

## Warning: Computation failed in `stat_bindot()`:
## attempt to apply non-function

## Warning: Computation failed in `stat_bindot()`:
## attempt to apply non-function

## Warning: Computation failed in `stat_bindot()`:
## attempt to apply non-function

## Warning: Computation failed in `stat_bindot()`:
## attempt to apply non-function

## Warning: Computation failed in `stat_bindot()`:
## attempt to apply non-function

## Warning: Computation failed in `stat_bindot()`:
## attempt to apply non-function

Now the plots live in things like wg1_violin_culture_comparisons$c0_violin_comparison

Cool now across I actually don’t think the other comparison makes sense because you’re comparing things across different gestures which defeats the purpose of the visualizations.

So anyway,

Stats

Could do ANOVAs but it doesn’t necessarily make sense since we’re just comparing two groups.

Now all that data lives in things like wg1_total_cultural_comparison_tables$c0_comparison that look like this: WG1-0 (this name isn’t here in the actual version, we just know from the naming scheme)

WG1

groups western_mean eastern_mean p sig
openness 3.710526 4.107143 1.00000
conflict 3.855263 4.160714 1.00000
control 4.210526 3.642857 0.69973
size 4.276316 4.178571 1.00000
certainty 4.921053 4.267857 0.13799
unity 5.105263 4.535714 0.71207

groups western_mean eastern_mean p sig
openness 3.900000 3.5000 1.00000
conflict 3.133333 3.9375 0.10881
control 4.416667 3.8125 0.29329
size 4.216667 3.8375 1.00000
certainty 5.166667 3.7750 0.00000 ***
unity 5.400000 4.2750 0.00039 ***

groups western_mean eastern_mean p sig
openness 2.986111 3.105263 1.00000
conflict 5.888889 4.986842 0.00703 *
control 4.694444 3.934210 0.02556 *
size 4.291667 4.197368 1.00000
certainty 4.625000 4.184210 1.00000
unity 3.625000 4.355263 0.20479

WG2

groups western_mean eastern_mean p sig
openness 3.381579 3.78 1.00000
conflict 5.486842 3.94 0.00001 ***
control 4.618421 3.48 0.00168 **
size 4.421053 4.12 1.00000
certainty 4.276316 4.54 1.00000
unity 4.263158 4.54 1.00000

groups western_mean eastern_mean p sig
openness 4.011905 2.937500 0.00014 ***
conflict 4.011905 4.921875 0.02176 *
control 3.892857 4.703125 0.03206 *
size 3.702381 3.406250 1.00000
certainty 4.797619 4.453125 1.00000
unity 4.702381 3.984375 0.03296 *

groups western_mean eastern_mean p sig
openness 3.657143 3.051282 0.21780
conflict 4.957143 4.474359 1.00000
control 4.600000 4.076923 0.69852
size 4.171429 3.564103 0.34673
certainty 4.957143 4.307692 0.06013 *
unity 4.928571 4.320513 0.22153

WG3

groups western_mean eastern_mean p sig
openness 3.756757 3.722222 1.00000
conflict 3.297297 3.722222 1.00000
control 4.324324 3.666667 0.26782
size 5.027027 4.962963 1.00000
certainty 4.783784 4.259259 0.23807
unity 5.445946 5.000000 0.79275

groups western_mean eastern_mean p sig
openness 3.684210 3.500000 1.00000
conflict 3.697368 3.833333 1.00000
control 4.881579 3.727273 0.00099 **
size 4.578947 4.106061 1.00000
certainty 4.500000 4.196970 1.00000
unity 4.907895 4.500000 1.00000

groups western_mean eastern_mean p sig
openness 4.739130 3.800000 0.00338 **
conflict 3.119565 3.950000 0.02874 *
control 3.923913 3.683333 1.00000
size 5.445652 4.833333 0.11437
certainty 4.989130 4.233333 0.00460 **
unity 5.554348 4.600000 0.00010 ***

WG4

groups western_mean eastern_mean p sig
openness 4.645161 4.7500 1.00000
conflict 3.145161 3.3250 1.00000
control 3.725807 3.1625 0.46166
size 5.209677 4.6000 0.16060
certainty 4.903226 4.8250 1.00000
unity 5.306452 4.9500 1.00000

groups western_mean eastern_mean p sig
openness 4.659091 4.158537 0.38857
conflict 3.545454 3.780488 1.00000
control 4.045454 3.536585 0.68870
size 5.431818 4.195122 0.00003 ***
certainty 4.954546 4.829268 1.00000
unity 4.954546 4.646342 1.00000

groups western_mean eastern_mean p sig
openness 4.111111 3.783784 1.00000
conflict 3.907407 4.067568 1.00000
control 4.462963 3.351351 0.00091 **
size 4.777778 4.527027 1.00000
certainty 4.444444 4.337838 1.00000
unity 4.740741 4.851351 1.00000

WG5

groups western_mean eastern_mean p sig
openness 4.454546 4.214286 1.00000
conflict 2.886364 3.273810 1.00000
control 3.318182 3.297619 1.00000
size 4.454546 3.857143 0.47394
certainty 4.363636 4.178571 1.00000
unity 5.159091 4.857143 1.00000

groups western_mean eastern_mean p sig
openness 4.947368 3.9625 0.01141 *
conflict 3.578947 3.2625 1.00000
control 4.157895 3.4875 0.56558
size 4.684210 3.7375 0.00638 *
certainty 5.078947 3.8125 0.00045 ***
unity 5.894737 4.3375 0.00000 ***

groups western_mean eastern_mean p sig
openness 4.270833 4.390625 1.00000
conflict 4.000000 3.578125 1.00000
control 3.916667 3.656250 1.00000
size 4.708333 4.171875 0.32606
certainty 4.479167 4.265625 1.00000
unity 4.875000 4.484375 1.00000

EG1

groups western_mean eastern_mean p sig
openness 3.939394 4.000000 1
conflict 3.893939 3.862069 1
control 3.636364 3.189655 1
size 4.303030 4.275862 1
certainty 4.212121 4.465517 1
unity 4.772727 4.862069 1

groups western_mean eastern_mean p sig
openness 3.367647 3.166667 1.00000
conflict 4.647059 4.187500 1.00000
control 4.911765 4.229167 0.43542
size 4.529412 3.770833 0.04422 *
certainty 4.617647 4.395833 1.00000
unity 4.691177 4.437500 1.00000

groups western_mean eastern_mean p sig
openness 4.231482 3.685185 0.56295
conflict 4.157407 4.185185 1.00000
control 3.712963 3.277778 1.00000
size 4.379630 4.444444 1.00000
certainty 3.944444 4.092593 1.00000
unity 4.768518 4.648148 1.00000

EG2

groups western_mean eastern_mean p sig
openness 3.369565 3.785714 1.00000
conflict 4.445652 3.964286 1.00000
control 4.391304 3.607143 0.61269
size 4.293478 4.214286 1.00000
certainty 4.445652 4.357143 1.00000
unity 4.684783 4.642857 1.00000

groups western_mean eastern_mean p sig
openness 3.978723 3.75000 1.00000
conflict 3.255319 3.96875 0.11945
control 3.712766 3.81250 1.00000
size 4.893617 4.00000 0.01229 *
certainty 4.723404 4.00000 0.01770 *
unity 5.308511 4.68750 0.24359

groups western_mean eastern_mean p sig
openness 3.441177 3.452381 1.00000
conflict 4.686274 4.309524 1.00000
control 4.205882 3.714286 1.00000
size 4.666667 4.095238 0.41235
certainty 4.098039 4.309524 1.00000
unity 4.274510 4.452381 1.00000

EG3

groups western_mean eastern_mean p sig
openness 3.707317 3.696429 1.00000
conflict 3.317073 4.053571 0.07923 *
control 4.036585 3.946429 1.00000
size 4.707317 4.571429 1.00000
certainty 4.963415 4.714286 1.00000
unity 5.475610 4.982143 0.47833

groups western_mean eastern_mean p sig
openness 3.987179 4.142857 1.0000
conflict 4.000000 3.476190 1.0000
control 3.794872 3.571429 1.0000
size 4.717949 4.000000 0.1541
certainty 4.679487 4.785714 1.0000
unity 5.282051 4.880952 1.0000

groups western_mean eastern_mean p sig
openness 3.988095 3.475 1.00000
conflict 3.178571 4.425 0.00254 **
control 3.904762 3.300 1.00000
size 4.750000 4.275 1.00000
certainty 4.785714 4.150 0.58306
unity 5.547619 4.875 0.22946

EG4

groups western_mean eastern_mean p sig
openness 4.206522 3.578947 0.23434
conflict 3.336956 4.157895 0.04647 *
control 4.184783 3.605263 0.97028
size 5.141304 4.315790 0.01705 *
certainty 4.804348 4.578947 1.00000
unity 5.260870 4.315790 0.01021 *

groups western_mean eastern_mean p sig
openness 4.1875 3.729167 0.78680
conflict 3.9000 3.937500 1.00000
control 4.1625 3.833333 1.00000
size 5.0875 4.208333 0.01834 *
certainty 4.6125 4.916667 1.00000
unity 4.9750 4.791667 1.00000

groups western_mean eastern_mean p sig
openness 3.935897 3.600 1.00000
conflict 4.333333 4.000 1.00000
control 4.730769 3.600 0.00316 **
size 4.987179 4.425 0.48959
certainty 4.717949 4.425 1.00000
unity 4.923077 4.400 0.71649

Are these powers good enough?

Quick spot check…

Short Answer: yes because we grouped so n = 2n lol. Interestingly, only for the significant differences do we see powers > 0.8. I mean not that interestingly cause like that’s how effect size works but still. Anyway, power is definitely high enough.

Example usage:

cond_power <- calculate_power_for_condition(wg2_total, "conflict", "original")

Then you get a variable called cond_power you can use to see the actual power through cond_power$power. In this case our power is 0.9946474